The total mortality from malignant melanoma continues to rise despite partially successful efforts at control of risk factors and at early diagnosis. Two stages of melanoma have been identified by the University of Pennsylvania's Pigmented Lesion Group in a model based on the analysis of stepwise tumor progression. In the first stage, radial growth phase, the lesion ay be invasive but is non-tumorigenic and metastasis does not occur. In the second stage, vertical growth phase, there is tumorigenic growth in the mesenchyme to form a mass lesion. Metastasis occurs with a probability that can be estimated on the basis of a multivariable prognostic mode. Growth in the mesenchyme appears to correlate with changes in cell surface adhesion molecular expression as judged by studies in cell lines, in experimental animals, and by limited univariate analysis of marker expression in frozen sections of human melanomas. A matrix adhesion subunit, beta3 integrin, two cell adhesion molecules, MUC18, the co-stimulatory molecule ICAM-1, an other integrins including alpha2beta1, alpha4beta1 and alpha5beta1, appear to be among the most highly expressed and/or specific tumor progression markers for melanoma. The Group has access to archival paraffin-embedded blocks and/or microscopic slides of the melanomas from more than 3000 prospectively registered and followed patients who have been entered into a clinical database. This unique pathology resource together with newly developed immunohistochemical methods will allow the group to test for the first time in a robust analysis the hypothesis that expression of adhesion molecules correlates with survival in cases of human melanoma evaluated in situ. The finding that a marker correlates with survival suggests that it represents a molecule of importance to the metastatic phenotype, which in turn may prompt the exploration of therapeutic possibilities. Further, such markers may be of value in the development of new prognostic models that will add to the accuracy of survival prediction for therapeutic decision-making and the planning of therapeutic trials. Finally, as clinicians remove increasing numbers of clinically problematical pigmented lesions, pathologists are faced with increasingly difficult diagnostic problems. Tumor progression antigens, if they can be studied in routine archival material, could provide the basis for useful diagnostic tests. This project represents a logical continuation and extension of our previous basic science work into the clinical arena, and it is well connected to the studies of adhesion molecules in the projects of M Herlyn and D Speicher, and to the evaluation of new antibodies that will be produced in these projects and in that of D Herlyn.